Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Pull latest release #1253

Merged
merged 153 commits into from
Jan 31, 2025
Merged

Pull latest release #1253

merged 153 commits into from
Jan 31, 2025

Conversation

luciansmith
Copy link

No description provided.

luciansmith and others added 30 commits April 19, 2024 09:44
(I hope.)

Also, re-enable test, and add new one for loading from scratch.

The savedState files are different on different OSes because the size of things is different (notably 'unsigned long').
The test seems to work in some contexts but not others.  Committing this version so we can test it on python on other platforms.
The trick turned out to be that on linux, loading loadBinary into jit->compiledModuleBinaryStream corrupted the compiledModuleBinaryStream.  Instead, we now convert to and from a string, and that seems to work.

At the same time, I changed the factory to return Jit* objects instead of unique_ptr objects, which also seems to work, though that might be a purely cosmetic change.
Also, bugs appeared only after loading twice, so check that.
Because this changes the save file format slightly.
* Move string conversion of compiledModuleBinaryStream to Jit class itself.
* Remove spurious endl.
* Initialize all object children in ModelGeneratorContext's default constructor.
* Clean up Jit objects in factory test.
* Describe new state_saving test and how to fix it when it inevitably breaks.
- it now runs on macos machines with Apple Silicon chips

- gcc is now updated in all linux platforms

- Host architecture is no longer provided in the matrix and is extracted internally

- artifacts names now provide more info about the machine and the compiler they were built on
Convert to the new GITHUB_OUTPUT instead.  Might not work with 'message'; could need to use 'echo' instead?
In theory, we'll need to update to node20 soon; this gets us partway there, and also sets up some python things.
Hopefully should work with node20?  Might need v4 for that.  Relative paths don't work, cf actions/upload-artifact#176
luciansmith and others added 29 commits September 18, 2024 17:28
(No llvm for it.)
Without this i am getting `'uint32_t' is not a member of 'std'`. See
also https://gcc.gnu.org/gcc-13/porting_to.html "Header dependency
changes".
…pdated (the issue with nleq2 is addressed in the new version of libroadrunner-deps)
…o that it considers the condition in which no steady state is reached.
…plugins)

- Identified and corrected memory leaks in using RRStringArrayPtr, RRCDataPtr, and RRVectorPtr.
- Implemented unique_ptr for RRVector.
…erent-results-on-mac-m1s

Fixing the issue with getting different results on mac m1s
Move to numpy 2 and Python 3.13
…ak_of_auto2000_plugin

fix: resolve memory leaks in auto2000 plugin
Update getFullJacobian, getReducedJacobian to handle reactions between species in different compartments of differing volumes. Affects getEigenValues() results.
* Use MacOS13
* Use latest dependencies.
* Fix installed swig 4.2 on ubuntu 24.
(But also update the savedState_windows.rr file, since I could update that one.)
@luciansmith luciansmith merged commit fcfd0dc into gh-pages Jan 31, 2025
27 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants